Home/Mechanical/Aerodynamic optimisation of a 150 m wind turbine

Static pressure contours on the front of the final 150 m rotor at 10 m/s — Surface pressure on the final rotor at 10 m/s, from the front.

Aerodynamic optimisation of a 150 m wind turbine

Honours thesis, supervised by Dr Ming Zhao · ANSYS Fluent · 2025

Wind farms space turbines 7 to 10 rotor diameters apart, because a turbine behind another one runs in its wake. My honours thesis asked how tightly they can be packed: how close a second turbine can sit before the lost power outweighs the extra density. For turbines inline, the answer is five diameters.

What went into it

Blade: A simple first blade to get reps in, then a twisted 150 m rotor scaled from the NREL 5 MW, with 67.6% more peak power
Mesh: Fluent Meshing with polyhedral cells, 6.65 million elements, after ANSYS Meshing failed on the twisted blade
Solver: Full-rotor transient CFD in ANSYS Fluent: 47 runs, 1,944 CPU hours. Steady state was tried to save time and dropped
Second turbine: Emulated from the first one's wake, when a true two-rotor model wouldn't fit the time left

SolidWorks model of the three-bladed rotor, isometric view — The final blade, isometric.

Side view of the hub and blade root showing the twisted root section — Side-on at the hub, showing the twisted root section.

One blade highlighted with its aerofoil sections along the span — Aerofoil sections along the span, DU inboard to NACA outboard.

Dimensioned chord distribution along a 75 m blade — Chord distribution along the 75 m blade, 5.43 m maximum to 1.69 m at the tip.

The flow domain with inlet and outlet planes, labelled 5D upstream, 5D lateral and 15D long — The enlarged domain, 15 diameters long.

The domain with measurement planes at 2D, 4D, 6D and 8D behind the rotor — Measurement planes at 2, 4, 6 and 8 diameters downstream.

Polyhedral mesh sliced through the final blade, fine prism layers at the surface growing into large cells — The final mesh, sliced at the blade: prism layers on the surface, polyhedra growing out from it.

The polyhedral mesh of the rotating zone around the rotor, sliced lengthwise, cells refining toward the blades — The polyhedral mesh in the rotating (MRF) zone around the rotor, sliced.

Fine polyhedral surface mesh on the blade — Surface mesh on the blade.

Chart of power coefficient against y plus for the Design 3 mesh study — Cp against y+ across the eight meshes tested at 10 m/s and 10 rpm.

Cp against tip speed ratio for Design 3, transient and steady state curves overlaid — Cp against tip speed ratio at 10 m/s, transient over steady state.

The result

The second turbine made 1,106 kW two diameters back, rising to 2,572 at eight, against 2,640 in clean air. Eight diameters gets the pair to 98.7% combined yield against 89.6% at five, but a row spaced at five fits 60% more turbines into the same length. That's why five diameters is the recommendation for turbines inline.

Static pressure contours on the rear face of the Design 3 rotor at 10 m/s — Surface pressure at 10 m/s, rear.

Shaft power against wind speed for Design 3 — Power curve of the final blade at a tip speed ratio of 7.07.

Velocity contour on a plane two diameters downstream at 10 m/s — Axial velocity, two diameters back.

Velocity contour four diameters downstream at 10 m/s — Four diameters.

Velocity contour six diameters downstream at 10 m/s — Six diameters.

Velocity contour eight diameters downstream at 10 m/s — Eight diameters, recovered.

Downstream shaft power against spacing with a logarithmic trendline — Downstream shaft power against spacing, with the fit.

Combined yield percentage against downstream distance — Combined yield of the pair against spacing.

What I took from it

I started the year with coursework-level CFD and finished with 47 simulations and about 500 hours of my own time against a nominal 300. Most of what I learned was when to trust a number: a coarse mesh read high, steady state read low, and nothing meant anything until it had been checked. Next I'd run a true two-rotor model and test a staggered layout. The wake never gets wider than about 1.34 diameters, so offsetting the second turbine sideways should get close to full yield, but that's still an assumption.

Message: Send me a message
LinkedIn: in/jeremy-alec-osborne
Based in: Sydney, NSW